Improving Lucky Imaging Photometry

For most ground-based optical and near-infrared telescopes today, seeing is the limiting factor in spatial resolution. Turbulence in the layers of the atmosphere distorts the wavefront of light coming from astronomical objects and effectively smears out the signal over a larger area in the focal plane, called the seeing disc. Even at the best observatory sites, located at several kilometres above sea level, the seeing rarely falls below 0.5" (arcsecond), and a seeing below 1" is considered to be good. For comparison a seeing of 1" corresponds to the resolution limit of a 10 cm telescope.

Adaptive optics systems can be used to mitigate the effects of the atmospheric turbulence, but these systems are very complicated and expensive and therefore not cost-effective for smaller telescopes. Another solution for this problem is the use of high frame-rate imaging. At very short exposure times (∼ 10 ms) it is possible to make snapshots of the wavefront before it is changed to a new configuration by the atmospheric turbulence. By analysing the speckle pattern in each snapshot, one may counteract the atmospheric disturbances by stacking the brightest speckles on top of each other and thus improve the spatial resolution. Even using longer exposure times (∼ 100 ms) this method can be used to mitigate the effect of image motion created by atmospheric turbulence.

The Lucky Imaging technique makes use of the fact that there is some probability that the wavefront on some of these snapshots has traverse the atmosphere almost unperturbed. If only these snapshots are stacked, it is possible to achieve very high spatial resolution.

Due to the readout noise it is not feasible to use conventional CCDs for high frame-rate imaging, unless bright sources are observed. In an electron multiplying CCD, or EMCCD, the signal is cascade amplified before it is read out, thus rendering the readout noise negligible compared to the signal. EMCCDs thus makes high frame-rate imaging feasible, even when observing faint objects.

In this thesis I describe the Two-Colour EMCCD Instrument (TCI) that has been designed for the SONG 1m telescopes and for the Danish 1.54m telescope. The motivation behind the TCI is to enable simultaneous red and visual band observation, thus providing instantaneous colour information and maximising the detected wavelength range without losing spatial information.

My main contribution has been developing the software for controlling the TCI and reducing the EMCCD data. This software includes, for instance, a fully automated routine for observing and processing microlensing events, such that light curves are automatically uploaded to a central server.

With the high spatial resolution that can be achieved with the TCI, it has, among other things, been possible to obtain high precision time-series photometry of crowded fields. I detail the discovery of two previously unknown variable stars in the Galactic globular cluster NGC 6981, and the search for variable stars in the five metal-rich ([Fe/H] > -1) globular clusters: NGC 6388, NGC 6441, NGC 6528, NGC 6638, NGC 6652.